370 1. Abstract 2. Introduction: Natural Killer Cells Kill

[Frontiers in Bioscience, Landmark, 22, 370-384, January 1, 2017]

Novel strategies to target cancer stem cells by NK cells; studies in humanized mice

Anna K. Kozlowska1,3, Kawaljit Kaur1, Paytsar Topchyan1, Anahid Jewett1,2

1Division of Oral Biology and Oral Medicine, The Jane and Jerry Weintraub Center for Reconstructive Biotechnology, UCLA, Los Angeles, CA, USA, 2The Jonsson Comprehensive Cancer Center, UCLA School of Dentistry and Medicine, Los Angeles, CA, USA, 3Department of Tumor Immunology, Chair of Medical Biotechnology, Poznan University of Medical Sciences, Poznan, Poland

TABLE OF CONTENTS

1. Abstract 2. Introduction: Natural killer cells kill and differentiate cancer stem-like cells 3. Studies of NK cells in xenogeneic implantation of human cells into immunodeficient mouse strains 4. Use of immunodeficient mouse strains in the studies of cancer immunity 5. Humanized mice as preclinical models to study the complexity of human immune system interactions 6. Potential limitations of allogeneic tumor transplantation in humanized mice 7. Are NK cells in humanized mice of sufficient quantity and quality? 8. NK cell receptor downregulation as potential mechanism for the detection of low NK cell frequencies in vivo 9. Adoptive therapy with osteoclast-expanded NK cells eliminates cancer stem-like cells in humanized mice 10. Future of NK cell mediated immunotherapy 11. Acknowledgements 12. References 1. ABSTRACT

We have previously shown, that following represent the first line of defense against virally infected selection, natural killer (NK) cells differentiate cancer stemcells and transformed cells. However, decreased NK like cells (CSCs)/poorly differentiated tumors via secreted cell cytotoxicity in the tumor microenvironment and and membrane bound IFN-gamma and TNF-alpha, peripheral blood of cancer patients, as well as down- leading to prevention of tumor growth and remodeling modulation of CD16 receptors on the surface of NK cells, of the tumor microenvironment. Since conventional have been reported and thought to contribute to cancer therapeutic strategies, including chemotherapy and progression (1, 2). Increasing evidence indicates that the radiotherapy remain unsuccessful in treating stem-like majority of human cancer originates and is maintained tumors, there has been increasing interest in NK cell- by proliferating stem-like populations, known as cancer targeted immunotherapy for the treatment of aggressive stem-like cells (CSCs), which are capable of self- malignacies. In our recent studies, we used a humanized renewal. Even though CSCs are highly susceptible to (hu-BLT) mouse model with transplanted human bone NK cells, they maintain resistance to chemotherapeutic marrow, liver and thymus to demonstrate the efficacy of drugs and radiation through increased expression of adoptive transfer of ex vivo expanded, super-charged NK DNA mismatch repair genes and augmented multi-drug cells in selection and differentiation of stem-like tumors resistance genes (3, 4). within the context of a fully reconstituted human immune system. We have also shown that CSCs differentiated with We have previously shown that NK cells split anergized NK cells prior to implantation in humanized cytotoxicity is suppressed after interaction with healthy mice did not grow or metastasize. In this review, we hematopoietic stem cells and malignant stem-like present current advances in NK cell detection, expansion cells (5-11). It has also been demonstrated that NK cells and therapeutic delivery methods, and discuss the utility lose the ability to mediate cytotoxicity and down-modulate of different humanized mouse models in studying NK cell- CD16 receptor expression upon interaction with CSCs, based therapies in the preclinical setting. while maintaining increased secretion of IFN-gamma, a functional outcome that we have previously coined as 2. INTRODUCTION: NATURAL KILLER split anergy (5-11). In contrast, differentiated tumor cells CELLS KILL AND DIFFERENTIATE CANCER do not down-modulate CD16 expression (7, 8). STEM-LIKE CELLS We have found that CSCs of different origins Natural killer (NK) cells constitute 10% to display low levels of MHC class I, CD54 and programmed 15% of human peripheral blood lymphocytes, and cell death protein ligand 1 (PD-L1), also known as

370 NK cells as therapy to eliminate CSCs in humanized mice

Figure 1. Hypothetical representation of osteoclast-expanded super-charged NK cell function in selection and differentiation of CSCs. Recognition of CSCs through CD16 and NKp46 NK cell receptors in a combination with Toll-like or cytokine receptors triggering on NK cells results in selection of CSCs subpopulation and differentiation of selected stem cells by split-anergized NK cells. CD56dimCD16bright NK cells are cytotoxic and are capable of selecting CSCs, whereas CD56dim/brightCD16dim split anergized NK cells lack cytotoxicity and are able to secrete high levels of cytokines (IFN-gamma and TNF-alpha) to support differentiation of selected stem cells. Please note, the receptors shown on NK cells are a partial list of receptors important for the activation of NK cell function. NK, NK cells; NK(SpA) cells, split anergized NK cells; TLR, toll-like receptors; S, stem cells; CD16R, CD16 receptors; NKp46R, NKp46 receptors; IL-2R, IL-2 receptors; DS, dead stem cells.

B7H1. A subset of CSCs become targets of cytotoxic translating data obtained from the mouse model to the NK cells, while the remaining CSCs are differentiated human system are structural and functional differences by split anergized/regulatory NK cells through cell- between human and murine immune systems. The cell contact and secretion of membrane bound and major cell population in human blood is represented by secreted forms of TNF-alpha and IFN-gamma (12). neutrophils (50-70%), whereas lymphocytes constitute In addition, differentiation by split anergized NK cells only 30-50%. Murine blood contains mainly lymphocytes leads to inhibition of CSCs growth and upregulation of (75-90%) and B cells are the most abundant type of MHC class I, CD54 and PD-L1 expression, resulting lymphocyte (18). Differences in human and mouse innate in resistance of differentiated tumor cells to NK cell- immune systems include the lack of defensin expression mediated cytotoxicity (12, 13). Overall, our studies by murine granulocytes, expression of different Toll- indicated that NK cells target and differentiate both like receptors on various immune subsets, and healthy and transformed stem-like cells, resulting in phenotypic differences in monocytes, DCs and MDSCs the maturation of target cells and shaping of their (reviewed in (18)). Despite similar proportions of NK microenvironment (Figure 1) (14-17). cells in murine and human blood, their phenotypes and activation dynamics are quite different (18, 19). CD3-/ Data obtained from in vitro cell culture studies CD16+CD56dim phenotype is typical for cytotoxic NK cells, often do not mirror the behavior of primary tumors in whereas the regulatory CD3-/CD16negCD56bright NK cells patients. Mice are chosen as the experimental model mediate differentiation of stem cells in humans (20, 21). in cancer immunology and preclinical testing of cancer Murine NK cells can be characterized by DX5 expression immunotherapeutics as they provide valuable insights and CD11b+/CD27- or CD11b-/CD27+ phenotypes, into the behavior of tumors in the context of full immune respectively (19, 22, 23). We have recently shown that cell repertoire. However, the major limitations of primary non-activated human NK cells mediate low-level

371 © 1996-2017 NK cells as therapy to eliminate CSCs in humanized mice cytotoxicity, but upon short-term IL-2 activation, they solid tumor engraftment compared to nude mice (31). display significant cytotoxicity against a number of Lapidot et al. demonstrated that low engraftment levels human stem-like tumors. In contrast, naïve murine NK of human bone marrow cell suspensions were present, cells do not display any cytotoxicity and require a much but as human cells represented only 0.5%-5% of the longer IL-2 activation period to induce cytotoxicity against total scid recipient marrow population (32), a more target cells (24). Human NK cells utilize KIR proteins as effective model was still needed. The low xenograft inhibitory receptors for MHC class I molecules, whereas reconstitution levels could not only be explained by the murine NK cells use highly divergent structures of Ly49 lack of critical human cytokines but also by the enhanced protein family members. In addition, human NKG2D function of the innate immune system. Despite the block receptors bind to MICA, MICB or ULBP 1-6 ligands, in lymphoid differentiation, C.B17-scid mice develop whereas mouse ligands for NKG2D are different and NK cells, as well as myeloid cells (33). It has been include, HAE60 and Rae1beta (18, 19). Therefore, these shown that, C.B17-scid NK cells could be stimulated differences should be considered and caution should with poly I: C to a higher extent than NK cells from wild be exercised when translating results of mouse tumor type mice, and displayed cytotoxicity against human models to human cancer. mesenchymal stem cells (MSCs), human embryonic stem cells (hESCs) and tumor initiating cells (TICs) In this review, we discuss the utility of humanized (34-36). This enhanced NK cell activity was reported mouse models as a preclinical platform to develop and in strains carrying scid mutations and was considered test novel NK cells based cancer immunotherapies. a compensation mechanism for the lack of adaptive Additionally, we discuss the use of osteoclast-expanded immunity. Similarly, we have recently shown that NK NK cells in the elimination of cancer stem-like tumors in cells purified from Cox-2flox/flox;LysMCre/+ mice have bone marrow, liver, thymus humanized (hu-BLT) mice. heightened cytotoxic activity when compared to those obtained from control littermates (24). In addition, NK 3. STUDIES OF NK CELLS IN XENOGENEIC cells cultured with autologous Cox-2flox/flox;LysMCre/+ IMPLANTATION OF HUMAN CELLS INTO monocytes, DCs and mouse embryonic fibroblasts IMMUNODEFICIENT MOUSE STRAINS (MEFs) isolated from global COX-2 knockout mice, had increased cytotoxic function as well as augmented There have been numerous attempts to IFN-gamma secretion when compared to NK cells establish an appropriate small animal model to study from control littermates cultured with monocytes. the complex interactions between immune cells and The list of genes that augment NK cell function when human tumors. Immunocompetent mouse strains knocked out in neighboring cells is still increasing, and are useless as hosts for human tumors because they may point to the fundamental function of NK cells in mount a xenogeneic immune response against human targeting less differentiated cells in order to aid in their cells, resulting in the failure of tumor engraftment. differentiation (13, 15, 24, 37). Several xenograft mouse models have previously been developed and are of high interest since they provide The rejection of tumor and hematopoietic information regarding the pathophysiology of human xenografts was partially alleviated by the development scid tumors in animals; however, the options to study the of the NOD-scid strain via introduction of the Prdkc interaction of human cancer cells with human immune mutated gene from C.B17-scid mice into a NOD system remain limited. As human HLA molecules are inbred strain with several impairments in innate not cross-reactive with murine NK cell receptors, in immunity (38). The resulting NOD-scid mice were the earliest xenograft models, certain human cells deficient in NK cells, displayed reduced development were likely recognized as MHC class I negative cells, and function of macrophages and dendritic cells and and were eliminated by murine NK cells (18, 25). The the absence of hemolytic complement (39, 40). In first immunodeficient strains used for xenogeneic contrast to scid mice, activity of NK cells remained low transplantation were T- cell deficient athymic nude in NOD-scid strain; thus pretreatment with poly I: C was (Foxn1) mice and severe combined immunodeficiency needed to detect active NK cells previously reported (scid) mice (26-28). The major disadvantages of nude as NK1.1. negative (39). NOD-scid mice expressed mice were their intact humoral adaptive immune system polymorphic variant of Sirp-alpha, which prevented + and high activity of murine NK cells, which prevented activation of murine macrophages against human CD47 most primary human solid tumors from seeding or cells (41, 42). Human hematopoietic stem cells express growing, and virtually disqualified engraftment of human the CD47 surface marker, which binds to Sirp-alpha normal or malignant hematopoietic cells (29). protein during engraftment in the bone marrow (42). High affinity binding of the Sirp-alpha to human CD47 The discovery of the spontaneous “scid” molecule prevents mouse macrophages from engulfing mutation in C.B17 strain (26), which destabilized mouse the human hematopoietic stem cells. Although NOD-scid hematopoietic stem cells (30) and prevented T and B mice support the growth of a large numbers of solid cells development enabled a broader range of human tumors and hematological malignancies, still a portion

372 © 1996-2017 NK cells as therapy to eliminate CSCs in humanized mice of tumors fails to engraft or grow efficiently, because of in their ability to support reconstitution of the human the remaining NK cell activity and other residual innate immune system, and since a profound immunodeficiency immune functions (34, 39). in the background strain is critical, the strain of choice can either be NSG or NRG (47, 48). There are three major The most recent introduction of a genetically humanized mouse models, which differ in the source engineered complete null mutation of the gamma chain and donor cell type, injection route and recipient age. of interleukin 2 receptor (IL2rg) into NOD-scid mice gave The most straightforward humanization model requires rise to one of the most immunodeficient mouse strains adoptive transfer of human peripheral blood mononuclear null known to date - NOD-scid IL2Rg (NSG) (43). IL2rg is cells (PBMCs) obtained from adult healthy donors or a common component of cell surface receptors for six patients into the NSG mice (49, 50). The advantages of different interleukins (IL-2, IL-4, IL-7, IL-9, IL-15, and such an approach are simplicity, immediate presence IL-21). Although, the cytokines are still present, IL2rg of mature cells circulating in recipients and the possibility knockout blocks molecular pathways for these cytokines, to study immune cells from patients in personalized resulting in the absence of functional NK cells since mouse systems. However, the utility of this model is they require IL-15 signaling for their development. The limited to short term experiments due to mature human remaining components of mouse innate immune cells immune cells initiating graft versus host (GvHD) disease include monocytes and neutrophils, as well as defective against their host (51). Another strategy involves prior macrophages and dendritic cells. NSG similarly to NOD- isolation of CD34+ progenitor cells from peripheral blood, null null Rag1 IL2Rgamma (NRG) strains are profoundly cord blood or fetal liver and injection into myeloablated immunodeficient, which is the key feature that supports NSG mice. This model requires 8-12 weeks as HSCs the growth of various malignancies, and most importantly, need to engraft bone marrow, differentiate and mature differentiation of human hematopoietic stem cells into into the various hematopoietic lineages of the human multi-lineage subsets (41, 44). immune system. After 8-12 weeks, mature human immune subsets can be easily detected in peripheral 4. USE OF IMMUNODEFICIENT MOUSE blood and other lymphoid tissues. The major limitation STRAINS IN THE STUDIES OF CANCER of the CD34+ humanization model is that due to the IMMUNITY lack of human thymus, T cells undergo selection in the context of the mouse histocompatibility complex (41). In Differences in the ability of CSCs to give rise to order to overcome this limitation, surgical implantation of human tumors in different immunodeficient strains could human fetal liver and thymus fragments under the renal be explained by different levels of NK cells impairment capsule of NSG mice is performed and followed by the and/or deletion in nude, NOD-scid and NSG strains (45). intravenous injection of autologous CD34+ hematopoietic As shown by Quintana et al., approximately 25% of cells to provide a complete human microenvironment for melanoma cells were able to initiate tumor growth in NSG immune cell development (52, 53). As a result, developing mice that were incapable of mediating any immune T cells undergo positive and negative selection in a function against injected stem-like cells, whereas only human thymic organoid and become functional CD4+ 0.1.% of melanoma cells could give rise to tumors in helper and CD8+ cytotoxic T cells after restriction NOD-scid animals with reduced but detectable NK cell based on HLA molecules in bone marrow, liver, thymus activity (46). humanized (hu-BLT) mice (41, 54). Hu-BLT is the only known humanized mouse model that displays mucosal Results of studies performed using immunity (55). HSCs develop, at least to some extent, immunodeficient animals have raised many questions into T, B, NK cells, monocytes, MDSCs, macrophages, about the role of certain immune subsets in controlling DCs, erythrocytes, and platelets in tissues of hu-BLT cancer initiation, growth and metastasis. As it is nearly mice (44, 56-58). Stable human CD45 cell engraftment impossible to assess and compare the aggressiveness in hu-BLT immune compartments is confirmed by flow and metastatic potential of primitive CSC-like tumors cytometry and contitutes approximately for 50% to 80% using immunodeficient mouse strains, it has been of cells in peripheral blood (manuscript in preparation). concluded that humanized mice with restored human Human immune cells are also found to populate the immune system would be the most suitable platform to reproductive tract of females, intestines, rectum (59, 60), conduct such studies. pancreas and gingiva (manuscript in prep). The broad use of hu-BLT in research may be limited by the availability 5. HUMANIZED MICE AS PRECLINICAL of fetal tissue and skilled surgical procedures. Moreover, MODELS TO STUDY THE COMPLEXITY OF since T cell education occurs in the context of human HUMAN IMMUNE SYSTEM INTERACTIONS thymus, T cells with affinity to murine MHC may still be present and as a result hu-BLT may develop GvHD-like Numerous attempts have been made to symptoms after 20 weeks post engraftment shortening generate mice with a fully reconstituted human immune the available experimental period (19). In contrast to system. Based on current evidence, mouse strains differ PBMC humanized mice, the hu-BLT model does not allow

373 © 1996-2017 NK cells as therapy to eliminate CSCs in humanized mice for personalized disease or genetic mutation study based in vitro observations, OSCSCs and pancreatic stem-like on cells obtained from patients. Despite such limitations, cancer cells differentiated with split anergized NK cell hu-BLT is currently the best available model for studying supernatants, grew much slower when injected at an human immunity. orthotopic site in hu-BLT mice ((12, 63) and manuscript in preparation). 6. POTENTIAL LIMITATIONS OF ALLOGENEIC TUMOR TRANSPLANTATION Similar data was obtained with melanoma IN HUMANIZED MICE cells in hu-BLT mice (manuscript in preparation). Our data is consistent with the observations from the JAX The humanized mouse model may be considered laboratories with Patient-Derived Xenografts (PDXs) + the best and closest mouse model to reflect the human of different origin in CD34 hu-mice without prior HLA immune system. The analysis of tumor-immune cell matching (64). However, because differentiation stages interactions within the tumor microenvironment including of the PDXs were not known, it is likely that the injected systemic tumor effects, as well as testing the efficacy tumors contained CSC populations capable of giving rise of anti-cancer therapies are attractive features of these to primary tumors in the presence of competent T cells. mice. The major presumed limitation of studies using Two different patient-derived melanoma cell lines have humanized mice appears to be the necessity to match been shown to be successfully implanted in hu-BLT HLA of the immune graft to the injected cancer cells. mice and treated with T cell based immunotherapy (65). Thus, the initial questions to be addressed were whether Melanoma tumors could only be cleared by MART-1 allogeneic tumors were likely to grow and if so, how transgenic T cells, HLA-matched with cancer cells, important HLA-matching was in mounting an effective whereas non-matched MART-1 T cells did not show any immune response in humanized mice. HLA mismatch therapeutic effect against melanoma. We have found between injected cancer cells and mature immune that these melanoma tumor cells had stem-like/poorly cells populating humanized mice has been expected to differentiated tumor cell phenotypes and were sensitive cause xenograft-allograft rejection or GvHD. Wege et al. targets for cytotoxic NK cells (unpublished data). proposed simultaneous co-engraftment of newborn NSG mice with CD34+ HSCs and human breast cancer cells Interestingly, in our studies, NSG immunodeficient to potentially avoid allograft rejection by mature immune mice developed larger tumors at a faster rate when cells (61). In such mice, human breast tumors were able to compared to hu-BLT mice. This result may account for the grow in the presence of maturating immune system, and severe lack of immune defense and control over tumor generation of specific T and NK cells responses as well growth in NSG mice (manuscript in preparation). Similarly, as infiltration of tumors with NK cells was observed (61). mice bearing K562 erythroleukemia tumors and injected with CD34+ HSC progenitors survived significantly longer Based on our previous data, we selected oral in comparison to their K562 bearing immunodeficient squamous cancer stem cells (OSCSCs) and pancreatic Balb/c Rag2-/- gammac -/- counterparts suggesting growth stem-like tumor cells as candidates to test tumor initiation inhibition of malignant cells by reconstituted human and growth in the presence of non-HLA matched mature immune cells (66). immune system in hu-BLT mice. We have previously shown that OSCSCs and pancreatic stem-like tumors We have previously demonstrated that expressed very low levels of MHC class I, CD54 and aggressive stem-like tumors, including melanoma, oral PD-L1, and they were highly susceptible to NK cell lysis, and pancreatic CSCs, are characterized by low classical whereas T cells were not able to target such tumors. MHC class I expression and could likely be resistant Furthermore, we demonstrated that oral and pancreatic to recognition and lysis by reconstituted or engrafted stem-like tumors differentiated upon interaction with autologous or allogeneic T cells which could, in part, split-anergized NK cells and their secreted factors, explain the lack of rejection of such cells by non-HLA- IFN-gamma and TNF-alpha (Figure 1). Differentiation matched T cells ((12, 13, 62), manuscript in preparation). with split-anergized NK cell supernatants increased On the other hand, cancer cells lacking MHC class I expression of MHC class I, CD54, and PD-L1 on cancer expression should be susceptible to lysis by NK cells. cells and rendered the tumors resistant to NK cell However, the data that we, and others, have obtained mediated cytotoxicity (14, 62). We demonstrated that suggest that autologous NK cells reconstituted in non-HLA matched OSCSCs and pancreatic stem-like humanized mice are less potent to effectively prevent cancer cells were able to form tumors in the oral cavity CSCs from establishing and growing in BLT mice. and pancreas of hu-BLT animals and were not rejected in the presence of competent T and B cells when 7. ARE NK CELLS IN HUMANIZED MICE OF injected after full reconstitution of the human immune SUFFICIENT QUANTITY AND QUALITY? system. We also observed infiltration of major human immune subsets including NK cells, T cells, B cells and Frequencies of human T cells and their subsets monocytes in the tumor microenvironment. Similar to circulating in humanized mice have been intensely

374 © 1996-2017 NK cells as therapy to eliminate CSCs in humanized mice investigated and proven to be similar to those of human of CD34+ NSG hu-mice, however, the number of CD56+ peripheral blood. T cells functionality was also tested, cells increased in all tissue compartments with plasmid however mostly in the context of HIV infection or other delivered IL-15, and the effect was more dramatic with infectious diseases (67-70). As education and selection combined treatment of IL-15 and Flt-3/Flk-2. However, of developing T cells in the context of human thymus is frequency of NK cells in blood decreased by 2-fold a critical for the generation of a broad repertoire of HLA- month after cytokine injection, suggesting the need for restricted T cells capable of mounting effective immune repeated delivery of IL-15 to maintain a functional pool of response, hu-BLT mice have become the model of choice NK cells (75). In contrast, humanized mice supplemented for HIV studies. with IL-7 were unable to develop functional NK cells and CD8+ T cells from hematopoietic cells isolated from Much less is known about the phenotype and juvenile patients with various malignancies (78). function of NK cells in humanized mice. In humans, studies on NK cells are mostly limited to those isolated Binding of IL-15 to IL2rg receptor requires trans- from peripheral blood and thus are less representative presentation in complex with IL-15Ralpha, of a different of NK cell repertoire present in human tissues. NK cells source than the target cell (85) and simultaneous are also found in secondary lymphoid organs, with expression of IL-15 and IL-15Ralpha, within the same cell CD56bright population being enriched, especially in lymph is needed for this process (86). In general, IL-15Ralpha nodes (71, 72). Given that over 40% of human lymphocytes protein is as ubiquitously expressed as its reported are found in secondary lymphoid organs, whereas only transcript expression, whereas IL-15 protein expression 2% circulate in peripheral blood, it can be assumed that is much more restricted. In mice, myeloid or DC cells are such CD56bright populations significantly contribute to NK shown to be the major sources of IL-15Ralpha. It is not cell-mediated innate responses in humans; therefore, a clear which cells support human NK cell homeostasis careful analysis of NK cell numbers and functions in all in vivo. Recently, DCs have been proposed to utilize immune compartments of humanized mice is the key to cross-presentation (87). Lucas et al. suggested that our understanding of NK function (71, 73). Accumulated human DCs might serve as a source of IL-15Ralpha, evidence suggests that NK cell numbers and functions in which prime and activate NK cells in lymph nodes or at both CD34+ and hu-BLT mouse models do not precisely sites of inflammation; such cytotoxic CD56dim/CD16bright reflect those observed in humans (74-78). Along with NK cells might then migrate to the periphery (88, 89). IL2rg others, we have also observed low frequencies of NK knockout mice have defects in IL-2 and IL-15 signaling. cells in hu-BLT mice based on the use of current NK Although, there is cross reactivity between mouse and detection markers. In addition, we observed decreased human IL-15, poor response of human immune cells cytotoxicity of NK cells in peripheral blood, bone to mouse IL-15 were reported and it was suggested to marrow, spleen and liver of humanized animals in account for extrinsically low reconstitution of NK cells in comparison to those obtained from human peripheral humanized mice (77). Huntington and colleagues found blood. Implantation of stem-like tumors further decreased similarly low levels of reconstituted NK cells in secondary cytotoxicity and cytokine secretion of reconstituted NK lymphoid organs of CD34+ reconstituted humanized cells (manuscript in preparation). mice that ranged from 0.3-1.5.% of human lymphocytes present in those tissues (77). Thus, reconstitution of NK NK cell development and maintenance cell levels in secondary lymphoid organs of humanized is regulated by a variety of factors including mice was much lower in comparison to the levels IL-15 (79-81). Decrease in IL-15 expression has observed in spleens and lymph nodes from healthy been shown to result in the absence of mature NK humans, which are 5% and 7-50%, respectively (71). cells in both mice and humans (82, 83). It has been As expected, NK cell development in bone marrow of + demonstrated that NK cells developed in CD34 Balb/c humanized mice was improved upon delivery of human -/- -/- Rag2 IL2Rgamma humanized mice in the absence IL-15/IL-15Ralpha. Interestingly, IL-15 in the absence of of IL-15, however, they were detected in much lower receptor was not as effective indicating the requirement numbers, mainly in the thymus and lymph nodes (74). for both the ligand and the receptor (77). These NK cells displayed CD56bright/CD16- phenotype similar to NK cells isolated from human lymph nodes. Although NK cells reconstituted from human In addition, similar to our previously defined split anergy HSCs in NSG mice could be detected in all blood- in NK cells, Ferlazzo et al., demonstrated secretion of perfused organs, at least one half displayed NKp46/ IFN-gamma in the absence of cytotoxicity by CD56bright/ CD56- phenotype and exhibited decreased function CD16- NK cells after 24-hour activation with IL-2 (84) (13). resembling human cord blood (CB) immature NK Upon IL-15 treatment increased numbers of NK cells cells (58). Similar findings reported by several groups in thymus, lymph nodes, spleen, bone marrow and might explain differential behavior of reconstituted liver could be observed (58, 74, 75). Chen et al. also NK cells in comparison to human NK cells obtained showed that human NK cells were detected at very low from blood, and collectively suggest the pre-activation percentages in blood, bone marrow, lung, spleen and liver requirement for reconstituted NK cells in humanized

375 © 1996-2017 NK cells as therapy to eliminate CSCs in humanized mice mice in order to obtain fully functional and potent NK cells could be detected in various tissue compartments cells that can mimic human adult peripheral blood NK including those in the tumor microenvironment whereas cells (61, 77, 90). Whether NKp46+/CD56- subsets are lower frequencies of NK cells were identified when cells immature NK populations or those that have interacted in isolated from the same tissues were stained with common vivo with the targets and subsequently down-modulated NK cell receptor antibodies (manuscript in preparation). CD56 surface receptors requires further investigation. Thus, such observations underscore the need for a novel or improved detection methods to determine the true 8. NK CELL RECEPTOR DOWNREGULATION frequencies of receptor-low NK cells in vivo. AS A POTENTIAL MECHANISM FOR THE DETECTION OF LOW NK CELL 9. ADOPTIVE THERAPY WITH OSTEOCLAST- FREQUENCIES IN VIVO EXPANDED NK CELLS ELIMINATES CANCER STEM-LIKE CELLS IN HUMANIZED MICE Interaction of NK cells with CSCs results in downregulation of CD16 receptor, as well as other NK cell Currently, adoptive NK cell therapy has been activating receptors, which usually correlates with cancer less efficient than T cell therapy in treating cancer patients progression (91-94). In this regard, we have also analyzed due to a number of limitations. With existing strategies, the frequencies of circulating CD3-/CD14-/CD19-/CD16-/ fewer numbers of potent NK cells can be expanded and CD56dim/neg NK cells in stage III and IV progressing used for cancer immunotherapy compared to T cells. melanoma patients, and found higher percentages of Cytokine and tumor activated NK cells neither survive long CD16 and CD56 receptor low/negative NK cells in these enough nor maintain their cytotoxic function to effectively patients in comparison to stage II melanoma patients eliminate tumors, due to rapid NK cell inactivation by the responding to therapy, as well as healthy individuals suppressive microenvironment in cancer patients. To (manuscript in preparation). The increased numbers circumvent such limitations we have recently established of such receptor low/negative NK cells in patients with a strategy to expand large numbers of functionally potent advanced stages of cancer might be the consequence super-charged NK cells using specific strains of sonicated of prior receptor signaling leading to down-modulation of probiotic bacteria in a combination with osteoclasts (95). key NK cell receptors by CSCs. As a result such NK cells This strategy has not only allowed for greater expansion become virtually undetectable if these receptors are used of activated NK cells accompanied by much higher levels for their detection. Similar receptor down-modulation may of cytotoxicity and cytokine secretion, but also prevented also occur for NK cells in the xenogeneic environment of NK cells from undergoing cell death when compared to humanized mice. It has been reported that reconstituted NK cells expanded by conventional strategies (manuscript NK cells display “anergized” phenotype that can be in preparation). We have previously shown that human characterized by downregulation of several receptors osteoclasts produced IL-15, IL-12, IL-18 and IFN-alpha, including CD16 (58, 75, 78). and displayed low expression of MHC class I and II, CD14, CD11b and CD54 (95). The use of osteoclasts The downregulation of receptors commonly as feeder cells preferentially expanded high numbers used for detection and phenotyping of NK cells raises of functionally potent NK cells that were maintained a question regarding the true frequencies of NK cells for a longer period of time in comparison to NK cells in cancer patients as well as in humanized mice. In our expanded by dendritic cells, monocytes, K562 cells and recent in vivo studies, we have observed that autologous CSCs. Even though dendritic cells were initially able to human NK cells isolated from hu-BLT mice with or expand NK cells, they could not maintain the expansion without tumors expressed very low levels of common and function of NK cells after the initial burst in NK cell NK cell receptors and displayed poor cytotoxicity. Similar activation. Dendritic cells preferentially expanded T cells, receptor down-modulation to those of autologous NK cells whereas osteoclasts specifically maintained expansion was also observed when osteoclast-expanded NK cells of NK cells with high cytotoxicity and cytokine secretion were injected into BLT humanized mice (manuscript in capabilities ((95) and manuscript in preparation). preparation). Interestingly, even though we could not see detectable numbers of NK cells using common NK cell Hu-BLT mice, which exhibit lower frequencies markers at the initiation of in vitro cultures, we detected of NK cells, serve as the best humanized mice for significant numbers of NK cells after a week of culture experimental NK cell therapy with ex vivo osteoclast- with IL-2. Further in vitro expansion and IL-2 activation of expanded human NK cells in human stem-like oral and such cells not only led to restoration of NK cell receptor pancreatic tumor models. Indeed, ex vivo osteoclast- expression but also augmented NK cell function. expanded human NK cells were effective in tumor cell lysis and limiting tumor growth in hu-BLT mice In order to track distribution of osteoclast- (manuscript in preparation). Significant reduction of tumor expanded NK in various tissue compartments, NK cells burden by expanded NK cells combined with whole cell were labeled with PKH-26 dye prior to intravenous cancer vaccine was also observed in the hu-BLT model injection to tumor bearing BLT mice. Red stained NK of advanced melanoma and the effect was accompanied

376 © 1996-2017 NK cells as therapy to eliminate CSCs in humanized mice by a dramatic change of cytokine secretion profiles by immunotherapeutic approaches in the context of the both immune and tumor cells (manuscript in preparation). reconstituted immune repertoire. Thus far, hu-BLT Interestingly, NK cells isolated from hu-BLT mice that animals provide the best available model for studying received osteoclast-expanded NK cells adoptive transfer the intricate human immune cell interactions with displayed higher cytotoxicity and proliferation rate in human tumors upon immunotherapeutics delivery, and comparison to autologous NK cells that were initially determining the molecular and cellular mechanisms of reconstituted in hu-BLT mice. tumor resistance leading to therapy failure. Moreover, this model can further be refined by the delivery of ex vivo It is possible that the maintenance of the derived, osteoclast-expanded, functionally potent human increased function of osteoclast-expanded, super-charged NK cells in order to overcome the observed limitations in NK cells after ex vivo re-stimulation from mice is a result of the numbers and function of autologous NK cells found prior cytokine activation and NK cell receptor binding and in humanized mice. signaling in the in vivo microenvironment, which imparts continued priming/signaling and the acquisition of memory- Our previous studies (12, 13, 97) and current like functions to NK cells. Even though NK cells may not in vivo studies indicate that NK cells are the main immune meet the criterion for classical memory that is reserved effectors that select and differentiate CSCs, resulting for T cells, it has been demonstrated that NK cells display in tumor growth inhibition and cessation of chronic higher function when reactivated after prior exposure to inflammation (13) (Figure 1). We have also demonstrated activation signals including cytokines or via engagement that NK cell-differentiated cancer cells become sensitive of activating NK receptors (96). It is known that freshly targets to chemotherapeutic and radiotherapeutic isolated “naïve” NK cells neither mediate cytotoxicity nor strategies (12) (Figure 1). In addition, our studies indicate spontaneously produce cytokines, however cytokine pre- that CSCs have, in general, low levels of MHC class I activated NK cells are capable of secreting IFN-gamma and therefore, are likely poor targets for T cells. NK cells upon transfer into naïve mice. Functions of such NK cells have the key role in differentiating such primitive cancer were reduced after a week following transfer but could be cells to provide differentiated targets for T cell-mediated restored after cytokine treatment or by the engagement effects. of their NK cell receptors, suggesting the existence of primed/memory-like NK cells (96). Our recent findings show great promise for the use of NK cells in cancer immunotherapy since both Importantly, since no side effects were observed autologous and allogeneic NK cells can be expanded after delivery of allogeneic NK cells within the xenogeneic ex vivo using osteoclasts, in order to not only substantially host microenvironment of hu-BLT mice, osteoclast- increase their numbers but also their functional potency expanded, super-charged NK cells may be safely used for successful treatment of cancer. Moreover, our data as a potential immunotherapy for human cancer patients. also indicates that the adoptive transfer of osteoclast- Indeed, we were able to optimize our expansion protocol expanded NK cells into hu-BLT mice initiates and in order to obtain high numbers of super-charged NK enhances recruitment of CD8+ T cells to the tumor cells required for repeated adoptive therapy of cancer microenvironment (Figure 1) (manuscript in prep). Thus, in humans by continuous expansion of NK cells for a adoptive transfer of super-charged NK cells may provide prolonged period of time using an optimized mixture of the basis for successful T cell based immunotherapies cytokines and receptor mediated signals. Contrary to all including tumor cell susceptibility to immune checkpoint previous expansion protocols used in clinical trials, we inhibitors and the use of peptides and cancer vaccines could continuously expand the same donor’s NK cells to boost immunity. Therefore, such combinatorial for an extended period of time with no functional loss. In addition, unlike cytokine-activated primary NK cells which approaches that target both CSCs as well as their more exhibited significant loss of numbers and function upon differentiated counterparts should result in better control freezing, osteoclast-expanded, super-charged NK cells of tumor growth, inhibit invasion, and prevent tumor effectively tolerated storage temperature and exhibited no immune escape and metastasis (Figure 1). significant reduction of cell viability or function after long term storage in liquid nitrogen. Such observations are 11. ACKNOWLEDGEMENTS critical for successful future immunotherapy, since patient- The authors declare no conflicts of interest. derived NK cells can be expanded with our method and Anna K. Kozlowska was supported by the Polish Ministry stored in liquid nitrogen for multiple future use. of Sciences and Higher Education and the Mobility Plus 10. FUTURE OF NK CELL MEDIATED award. IMMUNOTHERAPY 12. REFERENCES Based on our recent studies and those of others, we suggest that the hu-BLT mouse model is 1. P. Lai, H. Rabinowich, P. A. Crowley-Nowick, the best available preclinical model to study novel M. C. Bell, G. Mantovani and T. L. Whiteside:

Alterations in expression and function receptors mediates secretion of chemokines of signal-transducing proteins in tumor- and induces apoptotic death of naive natural associated T and natural killer cells in patients killer cells. Clin Cancer Res, 12(7 Pt 1), with ovarian carcinoma. Clin Cancer Res, 1994-2003 (2006) 2(1), 161-73 (1996) DOI: 10.1158/1078-0432.CCR-05-2306 2. I. Kuss, T. Saito, J. T. Johnson and T. L. 11. A. Jewett, A. Teruel, M. Romero, C. Head and Whiteside: Clinical significance of decreased N. Cacalano: Rapid and potent induction of cell zeta chain expression in peripheral blood death and loss of NK cell cytotoxicity against lymphocytes of patients with head and neck oral tumors by F(ab’)2 fragment of anti-CD16 cancer. Clin Cancer Res, 5(2), 329-34 (1999) antibody. Cancer Immunol Immunother, 57(7), 3. C. E. Meacham and S. J. Morrison: Tumour 1053-66 (2008) heterogeneity and cancer cell plasticity. DOI: 10.1007/s00262-007-0437-6 Nature, 501(7467), 328-37 (2013) 12. H.-C. Tseng, V. Bui, Y.-G. Man, N. Cacalano DOI: 10.1.038/nature12624 and A. Jewett: Induction of Split Anergy Conditions Natural Killer Cells to Promote 4. M. Cojoc, K. Mabert, M. H. Muders and A. Differentiation of Stem Cells through Cell- Dubrovska: A role for cancer stem cells in Cell Contact and Secreted Factors. Front therapy resistance: cellular and molecular Immunol, 5, 269-269 (2014) mechanisms. Semin Cancer Biol, 31, DOI: 10.3.389/fimmu.2014.0.0269 16-27 (2015) DOI: 10.1.016/j.semcancer.2014.0.6.0.04 13. H. C. Tseng, N. Cacalano and A. Jewett: Split DOI: 10.1016/j.semcancer.2014.06.004 anergized Natural Killer cells halt inflammation by inducing stem cell differentiation, resistance 5. A. Jewett, X. H. Gan, L. T. Lebow and B. to NK cell cytotoxicity and prevention Bonavida: Differential secretion of TNF-alpha of cytokine and chemokine secretion. and IFN-gamma by human peripheral blood- Oncotarget, 6(11), 8947-59 (2015) derived NK subsets and association with DOI: 10.18632/oncotarget.3250 functional maturation. J Clin Immunol, 16(1), 46-54 (1996) 14. H.-C. Tseng, V. T. Bui, Y.-G. Man, N. A. DOI: 10.1007/BF01540972 Cacalano and A. Jewett: Induction of split anergy conditions Natural Killer cells to 6. A. Jewett, M. Cavalcanti and B. Bonavida: promote differentiation of stem cells through Pivotal role of endogenous TNF-alpha in cell-cell contact and secreted factors. the induction of functional inactivation and Frontiers in immunology, 5, 269 (2014) apoptosis in NK cells. J Immunol, 159(10), DOI: 10.3.389/fimmu.2014.0.0269 4815-22 (1997) 15. A. Jewett, Y. G. Man and H. C. Tseng: Dual 7. A. Jewett and B. Bonavida: Target-induced functions of natural killer cells in selection and inactivation and cell death by apoptosis in a differentiation of stem cells; role in regulation subset of human NK cells. J Immunol, 156(3), of inflammation and regeneration of tissues. 907-15 (1996) J Cancer, 4(1), 12-24 (2013) 8. A. Jewett and B. Bonavida: Target-induced DOI: 10.7.150/jca.5519 anergy of natural killer cytotoxic function is 16. A. Jewett, H. C. Tseng, A. Arasteh, S. Saadat, restricted to the NK-target conjugate subset. R. E. Christensen and N. A. Cacalano: Natural Cell Immunol, 160(1), 91-7 (1995) killer cells preferentially target cancer stem DOI: 10.1016/0008-8749(95)80013-9 cells; role of monocytes in protection against 9. A. Jewett and B. Bonavida: MHC-Class I NK cell mediated lysis of cancer stem cells. antigens regulate both the function and the Curr Drug Deliv, 9(1), 5-16 (2012) survival of human peripheral blood NK cells: DOI: 10.2174/156720112798375989 role of endogenously secreted TNF-alpha. 17. A. Jewett and H. C. Tseng: Potential rescue, Clin Immunol, 96(1), 19-28 (2000) survival and differentiation of cancer stem cells DOI: 10.1006/clim.2000.4871 and primary non-transformed stem cells by 10. A. Jewett, N. A. Cacalano, C. Head and A. monocyte-induced split anergy in natural killer Teruel: Coengagement of CD16 and CD94 cells. Cancer immunology, immunotherapy:

CII, 61(2), 265-74 (2012) injected monocellular suspensions of human DOI: 10.1.007/s00262-011-1163-7 bone marrow to establish in the nude 18. J. Mestas and C. C. Hughes: Of mice and mouse. Int Arch Allergy Appl Immunol, 62(3), not men: differences between mouse and 330-3 (1980) human immunology. J Immunol, 172(5), DOI: 10.1159/000232530 2731-8 (2004) 26. G. C. Bosma, R. P. Custer and M. J. Bosma: DOI: 10.4049/jimmunol.172.5.2731 A severe combined immunodeficiency 19. A. Rongvaux, H. Takizawa, T. Strowig, T. mutation in the mouse. Nature, 301(5900), Willinger, E. E. Eynon, R. A. Flavell and M. G. 527-30 (1983) Manz: Human hemato-lymphoid system mice: DOI: 10.1038/301527a0 current use and future potential for medicine. 27. C. B. Isaacson JH: (Report). Mouse News Annu Rev Immunol, 31, 635-74 (2013) Lett., 27:31. (1962) DOI: 10.1.146/annurev-immunol-032712 28. E. M. Levy, D. Yonkosky, K. Schmid and S. -095921 R. Cooperband: Enrichment of the murine DOI: 10.1146/annurev-immunol- 032712 natural killer (NK) and mitogen induced cellular -095921 cytotoxicity (MICC) cells using preparative 20. A. Jewett, Y.-G. Man and H.-C. Tseng: Dual free-flow high voltage electrophoresis. Prep Functions of Natural Killer Cells in Selection Biochem, 7(6), 467-78 (1977) and Differentiation of Stem Cells; Role in DOI: 10.1.080/00327487708065514 Regulation of Inflammation and Regeneration 29. T. Nomura, T. Watanabe and S. Habu: of Tissues. Journal of Cancer, 4(1), 12-24 (2013) Humanized mice. Preface. Curr Top Microbiol DOI: 10.7.150/jca.5519 Immunol, 324, v-vi (2008) DOI: 10.1007/978-3-540-75647-7 21. M. A. Cooper, T. A. Fehniger and M. A. Caligiuri: The biology of human natural 30. G. Zhou, A. Hamik, L. Nayak, H. Tian, H. killer-cell subsets. Trends Immunol, 22(11), Shi, Y. Lu, N. Sharma, X. Liao, A. Hale, L. 633-40 (2001) Boerboom, R. E. Feaver, H. Gao, A. Desai, DOI: 10.1016/S1471-4906(01)02060-9 A. Schmaier, S. L. Gerson, Y. Wang, G. B. Atkins, B. R. Blackman, D. I. Simon and M. K. 22. W. M. Yokoyama, S. Kim and A. R. French: Jain: Endothelial Kruppel-like factor 4 protects The dynamic life of natural killer cells. Annu against atherothrombosis in mice. J Clin Rev Immunol, 22, 405-29 (2004) Invest, 122(12), 4727-31 (2012) DOI: 10.1.146/annurev.immunol.22.0.12703. DOI: 10.1.172/JCI66056 1.04711 DOI: 10.1146/annurev.immunol.22.012703. 31. R. A. Phillips, M. A. Jewett and B. L. Gallie: 104711 Growth of human tumors in immune-deficient scid mice and nude mice. Curr Top Microbiol 23. K. Yoshizawa, S. Nakajima, T. Notake, S. Immunol, 152, 259-63 (1989) Miyagawa, S. Hida and S. Taki: IL-15-high- DOI: 10.1007/978-3-642-74974-2_31 responder developing NK cells bearing Ly49 receptors in IL-15-/- mice. J Immunol, 187(10), 32. T. Lapidot, F. Pflumio, M. Doedens, B. Murdoch, 5162-9 (2011) D. E. Williams and J. E. Dick: Cytokine DOI: 10.4.049/jimmunol.1101561 stimulation of multilineage hematopoiesis from immature human cells engrafted in SCID 24. H. C. Tseng, A. Arasteh, K. Kaur, A. mice. Science, 255(5048), 1137-41 (1992) Kozlowska, P. Topchyan and A. Jewett: DOI: 10.1126/science.1372131 Differential Cytotoxicity but Augmented IFN- gamma Secretion by NK Cells after Interaction 33. K. Dorshkind, S. B. Pollack, M. J. Bosma with Monocytes from Humans, and Those and R. A. Phillips: Natural killer (NK) cells from Wild Type and Myeloid-Specific COX-2 are present in mice with severe combined Knockout Mice. Front Immunol, 6, 259 (2015) immunodeficiency (scid). J Immunol, 134(6), DOI: 10.3.389/fimmu.2015.0.0259 3798-801 (1985) 25. D. J. Ganick, R. D. Sarnwick, N. T. Shahidi 34. D. L. Greiner, R. A. Hesselton and L. D. Shultz: and D. D. Manning: Inability of intravenously SCID mouse models of human stem cell

engraftment. Stem Cells, 16(3), 166-77 (1998) J. E. Dick and J. S. Danska: Polymorphism DOI: 10.1.002/stem.160166 in Sirpa modulates engraftment of human 35. O. Livaditi, A. Kotanidou, A. Psarra, I. hematopoietic stem cells. Nat Immunol, 8(12), Dimopoulou, C. Sotiropoulou, K. Augustatou, 1313-23 (2007) C. Papasteriades, A. Armaganidis, C. DOI: 10.1.038/ni1527 Roussos, S. E. Orfanos and E. E. Douzinas: 43. L. D. Shultz, B. L. Lyons, L. M. Burzenski, Neutrophil CD64 expression and serum B. Gott, X. Chen, S. Chaleff, M. Kotb, S. D. IL-8: sensitive early markers of severity Gillies, M. King, J. Mangada, D. L. Greiner and outcome in sepsis. Cytokine, 36(5-6), and R. Handgretinger: Human lymphoid 283-90 (2006) and myeloid cell development in NOD/LtSz- DOI: 10.1.016/j.cyto.2007.0.2.0.07 scid IL2R gamma null mice engrafted with 36. X. Tian, P. S. Woll, J. K. Morris, J. L. mobilized human hemopoietic stem cells. Linehan and D. S. Kaufman: Hematopoietic J Immunol, 174(10), 6477-89 (2005) DOI: 10.4049/jimmunol.174.10.6477 engraftment of human embryonic stem cell- derived cells is regulated by recipient innate 44. F. Ishikawa, M. Yasukawa, B. Lyons, S. immunity. Stem Cells, 24(5), 1370-80 (2006) Yoshida, T. Miyamoto, G. Yoshimoto, T. DOI: 10.1.634/stemcells.2005-0340 Watanabe, K. Akashi, L. D. Shultz and M. Harada: Development of functional human 37. H. C. Tseng, V. Bui, Y. G. Man, N. Cacalano blood and immune systems in NOD/SCID/ and A. Jewett: Induction of Split Anergy IL2 receptor {gamma} chain(null) mice. Blood, Conditions Natural Killer Cells to Promote 106(5), 1565-73 (2005) Differentiation of Stem Cells through Cell- DOI: 10.1.182/blood-2005-02-0516 Cell Contact and Secreted Factors. Front Immunol, 5, 269 (2014) 45. L. D. Shultz, N. Goodwin, F. Ishikawa, V. DOI: 10.3.389/fimmu.2014.0.0269 Hosur, B. L. Lyons and D. L. Greiner: Human cancer growth and therapy in immunodeficient 38. S. Kataoka, J. Satoh, H. Fujiya, T. Toyota, R. mouse models. Cold Spring Harb Protoc, Suzuki, K. Itoh and K. Kumagai: Immunologic 2014(7), 694-708 (2014) aspects of the nonobese diabetic (NOD) DOI: 10.1.101/pdb.top073585 mouse. Abnormalities of cellular immunity. Diabetes, 32(3), 247-53 (1983) 46. E. Quintana, M. Shackleton, M. S. Sabel, DOI: 10.2337/diab.32.3.247 D. R. Fullen, T. M. Johnson and S. J. Morrison: Efficient tumour formation by single 39. L. D. Shultz, P. A. Schweitzer, S. W. human melanoma cells. Nature, 456(7222), Christianson, B. Gott, I. B. Schweitzer, B. 593-8 (2008) Tennent, S. McKenna, L. Mobraaten, T. DOI: 10.1.038/nature07567 V. Rajan, D. L. Greiner and et al.: Multiple defects in innate and adaptive immunologic 47. M. A. Brehm, A. Cuthbert, C. Yang, D. M. function in NOD/LtSz-scid mice. J Immunol, Miller, P. DiIorio, J. Laning, L. Burzenski, B. 154(1), 180-91 (1995) Gott, O. Foreman, A. Kavirayani, M. Herlihy, A. A. Rossini, L. D. Shultz and D. L. Greiner: 40. A. G. Baxter and A. Cooke: Complement lytic Parameters for establishing humanized activity has no role in the pathogenesis of mouse models to study human immunity: autoimmune diabetes in NOD mice. Diabetes, analysis of human hematopoietic stem cell 42(11), 1574-8 (1993) engraftment in three immunodeficient strains DOI: 10.2337/diab.42.11.1574 of mice bearing the IL2rgamma(null) mutation. 41. L. D. Shultz, M. A. Brehm, J. V. Garcia- Clin Immunol, 135(1), 84-98 (2010) Martinez and D. L. Greiner: Humanized mice DOI: 10.1.016/j.clim.2009.1.2.0.08 for immune system investigation: progress, 48. S. P. McDermott, K. Eppert, E. R. Lechman, promise and challenges. Nat Rev Immunol, M. Doedens and J. E. Dick: Comparison 12(11), 786-98 (2012) of human cord blood engraftment between DOI: 10.1.038/nri3311 immunocompromised mouse strains. Blood, 42. K. Takenaka, T. K. Prasolava, J. C. Wang, 116(2),193-200 (2010) S. M. Mortin-Toth, S. Khalouei, O. I. Gan, DOI: 10.1.182/blood-2010-02-271841

49. L. D. Shultz, F. Ishikawa and D. L. Greiner: Kosikova, J. M. Rivera, M. E. Moreno, B. Sloan, Humanized mice in translational biomedical P. Joshi and B. R. Long: Superior human research. Nat Rev Immunol, 7(2), leukocyte reconstitution and susceptibility 118-30 (2007) to vaginal HIV transmission in humanized DOI: 10.1.038/nri2017 NOD-scid IL-2Rgamma(-/-) (NSG) BLT mice. 50. A. Ito, T. Ishida, H. Yano, A. Inagaki, S. Suzuki, Virology, 417(1), 154-60 (2011) F. Sato, H. Takino, F. Mori, M. Ri, S. Kusumoto, DOI: 10.1.016/j.virol.2011.0.5.0.13 H. Komatsu, S. Iida, H. Inagaki and R. 56. E. Traggiai, L. Chicha, L. Mazzucchelli, L. Ueda: Defucosylated anti-CCR4 monoclonal Bronz, J. C. Piffaretti, A. Lanzavecchia and M. antibody exercises potent ADCC-mediated G. Manz: Development of a human adaptive antitumor effect in the novel tumor-bearing immune system in cord blood cell-transplanted humanized NOD/Shi-scid, IL-2Rgamma(null) mice. Science, 304(5667), 104-7 (2004) mouse model. Cancer Immunol Immunother, DOI: 10.1.126/science.1093933 58(8), 1195-206 (2009) 57. M. Ito, H. Hiramatsu, K. Kobayashi, K. DOI: 10.1.007/s00262-008-0632-0 Suzue, M. Kawahata, K. Hioki, Y. Ueyama, 51. M. A. King, L. Covassin, M. A. Brehm, W. Y. Koyanagi, K. Sugamura, K. Tsuji, T. Heike Racki, T. Pearson, J. Leif, J. Laning, W. and T. Nakahata: NOD/SCID/gamma(c)(null) Fodor, O. Foreman, L. Burzenski, T. H. mouse: an excellent recipient mouse model Chase, B. Gott, A. A. Rossini, R. Bortell, L. D. for engraftment of human cells. Blood, 100(9), Shultz and D. L. Greiner: Human peripheral 3175-82 (2002) blood leucocyte non-obese diabetic-severe DOI: 10.1.182/blood-2001-12-0207 combined immunodeficiency interleukin-2 58. T. Strowig, O. Chijioke, P. Carrega, F. Arrey, receptor gamma chain gene mouse model S. Meixlsperger, P. C. Rämer, G. Ferlazzo of xenogeneic graft-versus-host-like disease and C. Münz: Human NK cells of mice and the role of host major histocompatibility with reconstituted human immune system complex. Clin Exp Immunol, 157(1), components require preactivation to acquire 104-18 (2009) functional competence. Blood, 116(20), DOI: 10.1.111/j.1365-2249.2.009.0.3933.x 4158-67 (2010) 52. S. Shimizu, P. Hong, B. Arumugam, L. Pokomo, DOI: 10.1.182/blood-2010-02-270678 J. Boyer, N. Koizumi, P. Kittipongdaja, A. 59. R. Olesen, A. Wahl, P. W. Denton and J. V. Chen, G. Bristol, Z. Galic, J. A. Zack, O. Yang, Garcia: Immune reconstitution of the female I. S. Chen, B. Lee and D. S. An: A highly reproductive tract of humanized BLT mice and efficient short hairpin RNA potently down- their susceptibility to human immunodeficiency regulates CCR5 expression in systemic virus infection. J Reprod Immunol, 88(2), lymphoid organs in the hu-BLT mouse model. 195-203 (2011) Blood, 115(8), 1534-44 (2010) DOI: 10.1.016/j.jri.2010.1.1.0.05 DOI: 10.1.182/blood-2009-04-215855 60. P. W. Denton, R. Olesen, S. K. Choudhary, 53. D. N. Vatakis, G. C. Bristol, S. G. Kim, B. N. M. Archin, A. Wahl, M. D. Swanson, Levin, W. Liu, C. G. Radu, S. G. Kitchen and M. Chateau, T. Nochi, J. F. Krisko, R. A. J. A. Zack: Using the BLT humanized mouse Spagnuolo, D. M. Margolis and J. V. Garcia: as a stem cell based gene therapy tumor Generation of HIV latency in humanized BLT model. J Vis Exp(70), e4181 (2012) mice. J Virol, 86(1), 630-4 (2012) DOI: 10.3.791/4181 DOI: 10.1.128/JVI.06120-11 54. T. Onoe, H. Kalscheuer, N. Danzl, M. 61. A. K. Wege, W. Ernst, J. Eckl, B. Chittenden, G. Zhao, Y. G. Yang and M. Frankenberger, A. Vollmann-Zwerenz, D. Sykes: Human natural regulatory T cell N. Männel, O. Ortmann, A. Kroemer and G. development, suppressive function, and Brockhoff: Humanized tumor mice--a new postthymic maturation in a humanized mouse model to study and manipulate the immune model. J Immunol, 187(7), 3895-903 (2011) response in advanced cancer therapy. Int J DOI: 10.4.049/jimmunol.1100394 Cancer, 129(9), 2194-206 (2011) 55. C. A. Stoddart, E. Maidji, S. A. Galkina, G. DOI: 10.1.002/ijc.26159

62. H.-C. Tseng, A. Arasteh, A. Paranjpe, A. 69. L. Zhang and L. Su: HIV-1 immunopathogenesis Teruel, W. Yang, A. Behel, J. A. Alva, G. Walter, in humanized mouse models. Cell Mol C. Head, T.-o. Ishikawa, H. R. Herschman, Immunol, 9(3), 237-44 (2012) N. Cacalano, A. D. Pyle, N.-H. Park and A. DOI: 10.1.038/cmi.2012.7. Jewett: Increased Lysis of Stem Cells but Not 70. R. Akkina: Human immune responses Their Differentiated Cells by Natural Killer and potential for vaccine assessment in Cells; De-Differentiation or Reprogramming humanized mice. Curr Opin Immunol, 25(3), Activates NK Cells. Plos One, 5(7) (2010) 403-9 (2013) DOI: 10.1.371/journal.pone.0011590 DOI: 10.1.016/j.coi.2013.0.3.0.09 63. V. T. Bui, H. C. Tseng, A. Kozlowska, P. O. 71. G. Ferlazzo, D. Thomas, S. L. Lin, K. Maung, K. Kaur, P. Topchyan and A. Jewett: Goodman, B. Morandi, W. A. Muller, A. Augmented IFN-gamma and TNF-alpha Moretta and C. Munz: The abundant NK Induced by Probiotic Bacteria in NK Cells cells in human secondary lymphoid tissues Mediate Differentiation of Stem-Like Tumors require activation to express killer cell Ig-like Leading to Inhibition of Tumor Growth and receptors and become cytolytic. J Immunol, Reduction in Inflammatory Cytokine Release; 172(3), 1455-62 (2004) Regulation by IL-10. Front Immunol, 6, DOI: 10.4049/jimmunol.172.3.1455 576 (2015) DOI: 10.3.389/fimmu.2015.0.0576 72. S. Mizrahi, E. Yefenof, M. Gross, P. Attal, A. Ben Yaakov, D. Goldman-Wohl, B. 64. JAX Laboratory: Humanized NSG™ mice for Maly, N. Stern, G. Katz, R. Gazit, R. V. immuno-oncology (2015) Sionov, O. Mandelboim and S. Chaushu: 65. D. N. Vatakis, R. C. Koya, C. C. Nixon, L. A phenotypic and functional characterization Wei, S. G. Kim, P. Avancena, G. Bristol, D. of NK cells in adenoids. J Leukoc Biol, 82(5), Baltimore, D. B. Kohn, A. Ribas, C. G. Radu, 1095-105 (2007) Z. Galic and J. A. Zack: Antitumor activity DOI: 10.1.189/jlb.0407205 from antigen-specific CD8 T cells generated 73. J. Westermann and R. Pabst: Distribution of in vivo from genetically engineered human lymphocyte subsets and natural killer cells hematopoietic stem cells. Proc Natl Acad Sci in the human body. Clin Investig, 70(7), U S A, 108(51), E1408-16 (2011) 539-44 (1992) DOI: 10.1.073/pnas.1115050108 DOI: 10.1007/BF00184787 DOI: 10.1073/pnas.1115050108 74. E. A. Pek, T. Chan, S. Reid and A. A. Ashkar: 66. A. Kwant-Mitchell, E. A. Pek, K. L. Rosenthal Characterization and IL-15 dependence of and A. A. Ashkar: Development of functional NK cells in humanized mice. Immunobiology, human NK cells in an immunodeficient mouse 216(1-2), 218-24 (2011) model with the ability to provide protection DOI: 10.1.016/j.imbio.2010.0.4.0.08 against tumor challenge. PLoS One, 4(12), e8379 (2009) 75. Q. Chen, M. Khoury and J. Chen: Expression DOI: 10.1.371/journal.pone.0008379 of human cytokines dramatically improves DOI: 10.1371/journal.pone.0008379 reconstitution of specific human-blood lineage 67. C. C. Nixon, D. N. Vatakis, S. N. Reichelderfer, cells in humanized mice. Proc Natl Acad Sci U D. Dixit, S. G. Kim, C. H. Uittenbogaart and S A, 106(51), 21783-8 (2009) J. A. Zack: HIV-1 infection of hematopoietic DOI: 10.1.073/pnas.0912274106 progenitor cells in vivo in humanized mice. 76. M. G. Manz: Human-hemato-lymphoid- Blood, 122(13), 2195-204 (2013) system mice: opportunities and challenges. DOI: 10.1.182/blood-2013-04-496950 Immunity, 26(5), 537-41 (2007) 68. H. Garg, A. Joshi, C. Ye, P. Shankar and N. DOI: 10.1.016/j.immuni.2007.0.5.0.01 Manjunath: Single amino acid change in gp41 77. N. D. Huntington, N. Legrand, N. L. Alves, B. region of HIV-1 alters bystander apoptosis Jaron, K. Weijer, A. Plet, E. Corcuff, E. Mortier, and CD4 decline in humanized mice. Virol J, Y. Jacques, H. Spits and J. P. Di Santo: IL-15 8, 34 (2011) trans-presentation promotes human NK cell DOI: 10.1.186/1743-422X-8-34 development and differentiation in vivo. J Exp

Med, 206(1), 25-34 (2009) 101(47), 16606-11 (2004) DOI: 10.1.084/jem.20082013 DOI: 10.1.073/pnas.0407522101 78. M. C. Andre, A. Erbacher, C. Gille, V. 85. R. Koka, P. R. Burkett, M. Chien, S. Chai, Schmauke, B. Goecke, A. Hohberger, P. F. Chan, J. P. Lodolce, D. L. Boone and A. Mang, A. Wilhelm, I. Mueller, W. Herr, P. Ma: Interleukin (IL)-15R(alpha)-deficient Lang, R. Handgretinger and U. F. Hartwig: natural killer cells survive in normal but not Long-term human CD34+ stem cell-engrafted IL-15R(alpha)-deficient mice. J Exp Med, nonobese diabetic/SCID/IL-2R gamma(null) 197(8), 977-84 (2003) mice show impaired CD8+ T cell maintenance DOI: 10.1.084/jem.20021836 and a functional arrest of immature NK cells. 86. S. W. Stonier and K. S. Schluns: Trans- J Immunol, 185(5), 2710-20 (2010) presentation: a novel mechanism regulating DOI: 10.4.049/jimmunol.1000583 IL-15 delivery and responses. Immunol Lett, 79. B. Perussia, S. H. Chan, A. D’Andrea, K. 127(2), 85-92 (2010) Tsuji, D. Santoli, M. Pospisil, D. Young, S. DOI: 10.1.016/j.imlet.2009.0.9.0.09 F. Wolf and G. Trinchieri: Natural killer (NK) 87. L. Zhu, J. Butterton, A. Persson, M. Stonier, cell stimulatory factor or IL-12 has differential W. Comisar, D. Panebianco, S. Breidinger, effects on the proliferation of TCR-alpha beta+, J. Zhang and R. Bertz: Pharmacokinetics TCR-gamma delta+ T lymphocytes, and NK and safety of twice-daily atazanavir 300 mg cells. J Immunol, 149(11), 3495-502 (1992) and raltegravir 400 mg in healthy individuals. 80. E. Mrozek, P. Anderson and M. A. Caligiuri: Antivir Ther, 15(8), 1107-14 (2010) Role of interleukin-15 in the development of DOI: 10.3.851/IMP1673 human CD56+ natural killer cells from CD34+ hematopoietic progenitor cells. Blood, 87(7), 88. M. Lucas, W. Schachterle, K. Oberle, P. 2632-40 (1996) Aichele and A. Diefenbach: Dendritic cells prime natural killer cells by trans-presenting 81. T. Sato, J. H. Laver, Y. Aiba and M. Ogawa: interleukin 15. Immunity, 26(4), 503-17 (2007) NK cell colony formation from human DOI: 10.1.016/j.immuni.2007.0.3.0.06 fetal thymocytes. Exp Hematol, 27(4), 726-33 (1999) 89. M. A. Cooper, T. A. Fehniger, A. Fuchs, M. DOI: 10.1016/S0301-472X(99)00005-3 Colonna and M. A. Caligiuri: NK cell and DC interactions. Trends Immunol, 25(1), 82. C. A. Vosshenrich, T. Ranson, S. I. Samson, E. 47-52 (2004) Corcuff, F. Colucci, E. E. Rosmaraki and J. P. DOI: 10.1016/j.it.2003.10.012 Di Santo: Roles for common cytokine receptor gamma-chain-dependent cytokines in the 90. T. Strowig, O. Chijioke, P. Carrega, F. Arrey, generation, differentiation, and maturation of S. Meixlsperger, P. C. Ramer, G. Ferlazzo NK cell precursors and peripheral NK cells in and C. Munz: Human NK cells of mice vivo. J Immunol, 174(3), 1213-21 (2005) with reconstituted human immune system DOI: 10.4049/jimmunol.174.3.1213 components require preactivation to acquire functional competence. Blood, 116(20), 83. A. Fischer, F. Le Deist, S. Hacein-Bey-Abina, 4158-67 (2010) I. Andre-Schmutz, S. Basile Gde, J. P. de DOI: 10.1.182/blood-2010-02-270678 Villartay and M. Cavazzana-Calvo: Severe combined immunodeficiency. A model disease 91. G. Pietra, C. Manzini, S. Rivara, M. Vitale, for molecular immunology and therapy. C. Cantoni, A. Petretto, M. Balsamo, R. Immunol Rev, 203, 98-109 (2005) Conte, R. Benelli, S. Minghelli, N. Solari, M. DOI: 10.1.111/j.0105-2896.2.005.0.0223.x Gualco, P. Queirolo, L. Moretta and M. C. DOI: 10.1111/j.0105-2896.2005.00223.x Mingari: Melanoma cells inhibit natural killer 84. G. Ferlazzo, M. Pack, D. Thomas, C. Paludan, cell function by modulating the expression D. Schmid, T. Strowig, G. Bougras, W. A. of activating receptors and cytolytic activity. Muller, L. Moretta and C. Munz: Distinct roles Cancer Res, 72(6), 1407-15 (2012) of IL-12 and IL-15 in human natural killer cell DOI: 10.1.158/0008-5472.CAN-11-2544 activation by dendritic cells from secondary 92. T. Garcia-Iglesias, A. Del Toro-Arreola, B. lymphoid organs. Proc Natl Acad Sci U S A, Albarran-Somoza, S. Del Toro-Arreola, P. E.

Sanchez-Hernandez, M. G. Ramirez-Duenas, Send correspondence to: Anahid Jewett, 10833 L. M. Balderas-Pena, A. Bravo-Cuellar, P. Le Conte Ave, UCLA School of Dentistry, Los C. Ortiz-Lazareno and A. Daneri-Navarro: Angeles, CA 90095, USA, Tel: 310-206-3970, Fax: Low NKp30, NKp46 and NKG2D expression 310-794-7109, E-mail: [email protected] and reduced cytotoxic activity on NK cells in cervical cancer and precursor lesions. BMC Cancer, 9, 186 (2009) DOI: 10.1.186/1471-2407-9-186 93. Y. P. Peng, J. J. Zhang, W. B. Liang, M. Tu, Z. P. Lu, J. S. Wei, K. R. Jiang, W. T. Gao, J. L. Wu, Z. K. Xu, Y. Miao and Y. Zhu: Elevation of MMP-9 and IDO induced by pancreatic cancer cells mediates natural killer cell dysfunction. BMC Cancer, 14, 738 (2014) DOI: 10.1.186/1471-2407-14-738 94. E. Mamessier, A. Sylvain, F. Bertucci, R. Castellano, P. Finetti, G. Houvenaeghel, E. Charaffe-Jaufret, D. Birnbaum, A. Moretta and D. Olive: Human breast tumor cells induce self-tolerance mechanisms to avoid NKG2D-mediated and DNAM-mediated NK cell recognition. Cancer Res, 71(21), 6621-32 (2011) DOI: 10.1.158/0008-5472.CAN-11-0792 95. H. C. Tseng, K. Kanayama, K. Kaur, S. H. Park, S. Park, A. Kozlowska, S. Sun, C. E. McKenna, I. Nishimura and A. Jewett: Bisphosphonate- induced differential modulation of immune cell function in gingiva and bone marrow in vivo: Role in osteoclast-mediated NK cell activation. Oncotarget, 6(24), 20002-25 (2015) DOI: 10.18632/oncotarget.4755 96. M. A. Cooper, J. M. Elliott, P. A. Keyel, L. Yang, J. A. Carrero and W. M. Yokoyama: Cytokine- induced memory-like natural killer cells. Proc Natl Acad Sci U S A, 106(6), 1915-9 (2009) DOI: 10.1.073/pnas.0813192106 97. H. C. Tseng, A. Arasteh, A. Paranjpe, A. Teruel, W. Yang, A. Behel, J. A. Alva, G. Walter, C. Head, T. O. Ishikawa, H. R. Herschman, N. Cacalano, A. D. Pyle, N. H. Park and A. Jewett: Increased lysis of stem cells but not their differentiated cells by natural killer cells; de-differentiation or reprogramming activates NK cells. PLoS One, 5(7), e11590 (2010) DOI: 10.1.371/journal.pone.0011590 DOI: 10.1371/journal.pone.0011590

Key Words: Osteoclast-expanded Super Charged NK Cells, Cancer Immunotherapy, BLT Humanized Mice, CSCs, Review